We propose to investigate the process of chest lesion detection by selectively modifying the image quality of radiographs containing chest lesions. Our procedure consists of forming a lesion image mask and combining it with a normal radiograph to provide a test radiograph by means of a duplication process. Base line data on a group of observers have been collected in a pilot project. We now propose to modify the image quality of the test radiographs in a quantitative manner and to collect data on observer accuracy, sensitivity, specificity and ROC curves obtained with the test series having modified image quality. We will make comparisons with the base line data and will obtain measurements of the MTF of the modified image systems to attempt to relate physical changes in this parameter of image quality to observer response. Image quality will be modified by blurring during duplication (either the entire image or only the lesion) and producing changes in optical density, lesion contrast, edge sharpness, and observer's vision. We have also conducted pilot studies of other effects which could be expected to affect diagnostic accuracy such as observer fatigue and the value of providing a second image for side-by-side comparison. The pilot studies have demonstrated the validity of the use of duplicate radiographs for the test series and the ability to obtain observers for the test series. We plan follow-up studies on these pilot investigations with emphasis on clinical applications.